INDEX

Explanations

references to military technology and weaponry

oai_token-act-pair · gpt-4o-mini Triggered by @bot

quantities and measures related to weapons and military force

oai_token-act-pair · gpt-4o-mini Triggered by @bot

attends to weapon-related tokens (like "firearms", "weapons", "guns", "arms") from subsequent prepositions or conjunctions (like "of", "and", "or").

oai_attention-head · claude-3-5-sonnet-20240620 Triggered by @danielepoterti

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.9.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.04

1:0.02

2:0.12

3:0.13

4:0.03

5:0.09

6:0.03

7:0.28

8:0.12

9:0.02

10:0.06

11:0.01

Negative Logits

idth

-2.28

{{

-2.20

athom

-2.18

oward

-2.13

 Shoals

-2.10

antha

-2.08

 Peak

-2.06

icip

-2.06

+/-

-2.02

 Happ

-1.98

POSITIVE LOGITS

 ammunition

2.76

 grenades

2.74

IED

2.67

Medic

2.66

 grenade

2.66

 camouflage

2.59

 medic

2.55

HUD

2.55

 ammo

2.46

azine

2.42

Activations Density 0.046%

references to military technology and weaponry

quantities and measures related to weapons and military force

attends to weapon-related tokens (like "firearms", "weapons", "guns", "arms") from subsequent prepositions or conjunctions (like "of", "and", "or").

No Comments

No Known Activations