INDEX

Explanations

references to legal charges, specifically felonies

oai_token-act-pair · gpt-3.5-turbo

mentions of felony charges in legal contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arity

-0.91

tz

-0.83

zsche

-0.78

andr

-0.77

TPPStreamerBot

-0.75

arin

-0.74

choes

-0.74

obyl

-0.71

reen

-0.71

atu

-0.71

POSITIVE LOGITS

 convictions

0.93

 felony

0.92

 punishable

0.89

 offenses

0.87

 probation

0.87

 unfocusedRange

0.86

 perjury

0.86

 conviction

0.85

 criminal

0.85

 offences

0.85

Activations Density 0.009%

references to legal charges, specifically felonies

mentions of felony charges in legal contexts

No Comments

No Known Activations