INDEX

Explanations

mentions of legal indictments

oai_token-act-pair · gpt-3.5-turbo

the term "indictment" and its variations, indicating legal accusations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

atomic

-0.70

 Kers

-0.67

olen

-0.66

 grass

-0.65

birth

-0.65

********************************

-0.65

gm

-0.65

events

-0.64

 leisure

-0.64

perm

-0.64

POSITIVE LOGITS

 indictment

1.51

eering

0.99

 indict

0.95

 indicted

0.90

mination

0.75

Charges

0.73

 affidavit

0.71

FontSize

0.71

uary

0.71

gdala

0.69

Activations Density 0.006%

mentions of legal indictments

the term "indictment" and its variations, indicating legal accusations

No Comments

No Known Activations