INDEX

Explanations

mentions of legal or court-related terms and activities

oai_token-act-pair · gpt-3.5-turbo

references to legal proceedings and court-related terms

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

TRUMP

-0.71

Song

-0.69

kun

-0.68

travel

-0.67

âĺħ

-0.66

oute

-0.66

CDC

-0.66

LAN

-0.65

ãĤ¦

-0.65

wcs

-0.64

POSITIVE LOGITS

 adjourn

0.92

 magistrate

0.90

 sentencing

0.84

 hearings

0.82

 commissioners

0.80

 proceedings

0.77

 arra

0.76

 judge

0.73

istrate

0.73

 courtroom

0.73

Activations Density 0.094%

mentions of legal or court-related terms and activities

references to legal proceedings and court-related terms

No Comments

No Known Activations