INDEX

Explanations

mentions of discussing or dialoguing

oai_token-act-pair · gpt-3.5-turbo

phrases that include the word "talking."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iver

-0.73

hement

-0.67

cffff

-0.66

ridge

-0.64

hews

-0.64

edu

-0.64

falls

-0.63

ursor

-0.62

imon

-0.60

emale

-0.59

POSITIVE LOGITS

 about

1.16

 ABOUT

0.97

 bout

0.89

about

0.85

 About

0.78

 here

0.75

 figur

0.75

 specifically

0.72

 purely

0.68

About

0.66

Activations Density 0.070%

mentions of discussing or dialoguing

phrases that include the word "talking."

No Comments

No Known Activations