INDEX

Explanations

references to investigations or inquiries, particularly those involving scrutiny or questioning

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

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.03

2:0.13

3:0.06

4:0.10

5:0.07

6:0.03

7:0.03

8:0.13

9:0.20

10:0.06

11:0.03

Negative Logits

weet

-1.39

ahime

-1.25

natureconservancy

-1.22

olson

-1.19

alli

-1.17

uphem

-1.13

izabeth

-1.09

undle

-1.09

eva

-1.07

 appl

-1.07

POSITIVE LOGITS

 probing

1.55

 Probe

1.38

 probes

1.24

 probe

1.24

 Collider

1.21

 uncover

1.14

Leaks

1.10

��

1.07

NRS

1.07

 Plat

1.06

Activations Density 0.003%

references to investigations or inquiries, particularly those involving scrutiny or questioning

No Comments

No Known Activations