INDEX

Explanations

phrases indicating disassociation or denial of involvement

oai_token-act-pair · gpt-3.5-turbo

phrases that refute associations or claims by emphasizing a lack of connection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Dane

-0.68

cember

-0.65

 pregnancies

-0.61

 Gust

-0.59

ORN

-0.58

Pix

-0.58

rs

-0.57

 Drops

-0.56

 Controls

-0.56

riger

-0.56

POSITIVE LOGITS

 satisfy

0.98

 celebrate

0.93

 settle

0.93

 prove

0.90

 learn

0.89

 stop

0.87

 clarify

0.85

 defend

0.84

 remind

0.83

 give

0.82

Activations Density 0.079%

phrases indicating disassociation or denial of involvement

phrases that refute associations or claims by emphasizing a lack of connection

No Comments

No Known Activations