INDEX

Explanations

words related to accusations or claims of wrongdoing

oai_token-act-pair · gpt-3.5-turbo

instances of the word "alleged" and context related to allegations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ascript

-0.74

haar

-0.74

ciating

-0.73

ARCH

-0.72

eday

-0.72

patch

-0.71

aws

-0.70

eyes

-0.70

Ĺ

-0.69

arton

-0.69

POSITIVE LOGITS

 wrongdoing

1.08

 violations

0.93

 misconduct

0.90

 improper

0.89

 abuses

0.89

 indecent

0.88

 misuse

0.88

 misrepresent

0.88

 abuse

0.86

 wrongful

0.86

Activations Density 0.048%

words related to accusations or claims of wrongdoing

instances of the word "alleged" and context related to allegations

No Comments

No Known Activations