INDEX

Explanations

keywords related to accusations or allegations of wrongdoing

oai_token-act-pair · gpt-3.5-turbo

instances of the word "allegedly" related to various criminal activities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ERAL

-0.71

anim

-0.71

lite

-0.66

ABLE

-0.64

Shadow

-0.64

ilation

-0.63

Appearances

-0.63

egg

-0.62

then

-0.61

fork

-0.61

POSITIVE LOGITS

 violated

0.77

 plotted

0.75

 interfered

0.71

 violate

0.71

 infringing

0.71

 obtained

0.71

 threaten

0.71

 infring

0.71

 duct

0.70

 aggravated

0.69

Activations Density 0.019%

keywords related to accusations or allegations of wrongdoing

instances of the word "allegedly" related to various criminal activities

No Comments

No Known Activations