INDEX

Explanations

information related to political decisions and legal proceedings

oai_token-act-pair · gpt-3.5-turbo

phrases related to legislative actions or proceedings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sacrific

-0.63

tou

-0.60

 throb

-0.59

;)

-0.59

 lovely

-0.58

 sprite

-0.57

etc

-0.57

?).

-0.57

 mould

-0.56

!".

-0.56

POSITIVE LOGITS

ortium

0.85

 federal

0.77

enforcement

0.76

 issuance

0.74

 Fiscal

0.73

 preliminary

0.73

 Judicial

0.73

displayText

0.73

 subpoen

0.72

 redacted

0.71

Activations Density 1.774%

information related to political decisions and legal proceedings

phrases related to legislative actions or proceedings

No Comments

No Known Activations