INDEX

Explanations

verbs ending in 'ing', specific violence-related words, and words related to conflict and war

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

terms related to mass production and significant actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_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.3.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

0:0.02

1:0.02

2:0.06

3:0.35

4:0.01

5:0.02

6:0.12

7:0.10

8:0.03

9:0.06

10:0.05

11:0.10

Negative Logits

��

-1.34

itely

-1.24

ighth

-1.23

trak

-1.23

ogh

-1.22

arkin

-1.11

��

-1.10

ffer

-1.09

 Centauri

-1.09

 Seym

-1.08

POSITIVE LOGITS

 incarcer

1.27

ゼウス

1.14

hesive

1.12

arcer

1.11

 deport

1.10

ocese

1.08

endars

1.04

uttering

1.04

 bond

1.03

rier

1.02

Activations Density 0.008%

verbs ending in 'ing', specific violence-related words, and words related to conflict and war

terms related to mass production and significant actions or events

No Comments

No Known Activations

verbs ending in 'ing', specific violence-related words, and words related to conflict and war

terms related to mass production and significant actions or events

No Comments

No Known Activations