INDEX

Explanations

words related to legal outcomes or decisions

oai_token-act-pair · gpt-3.5-turbo

references to judicial decisions or verdicts

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

href

-0.66

 Heritage

-0.66

 Occupations

-0.64

STD

-0.62

ordinate

-0.62

cession

-0.60

ohyd

-0.60

ople

-0.60

avy

-0.59

Dak

-0.58

POSITIVE LOGITS

 verdict

1.11

ysis

1.04

naire

0.84

 Judgment

0.84

iciary

0.79

0.77

atorial

0.75

INAL

0.74

 debtor

0.73

ittal

0.73

Activations Density 0.009%

words related to legal outcomes or decisions

references to judicial decisions or verdicts

No Comments

No Known Activations