INDEX

Explanations

references to legal and criminal activities

oai_token-act-pair · gpt-3.5-turbo

phrases related to crime and legal issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

DK

-0.59

 Saga

-0.59

 Warlock

-0.56

 Kuro

-0.55

 Elves

-0.55

Yug

-0.52

ãĢį

-0.52

EVs

-0.51

 Innovation

-0.51

rying

-0.50

POSITIVE LOGITS

 forensic

0.67

 guiActiveUn

0.66

nsic

0.65

 appellant

0.64

 Forensic

0.63

 remorse

0.63

 criminal

0.63

hement

0.62

iazep

0.62

 jail

0.61

Activations Density 2.037%

references to legal and criminal activities

phrases related to crime and legal issues

No Comments

No Known Activations