INDEX

Explanations

phrases indicating communication or contact

oai_token-act-pair · gpt-3.5-turbo

instances of communication or messages from various sources

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Saharan

-0.73

 Athletic

-0.73

tein

-0.72

bis

-0.67

ritional

-0.67

ijn

-0.65

 Enix

-0.61

adelphia

-0.61

taboola

-0.59

 Homo

-0.57

POSITIVE LOGITS

 firsthand

0.89

 podcasts

0.87

 loud

0.79

 aloud

0.74

maid

0.73

 Voices

0.73

heard

0.72

 voices

0.68

SAY

0.68

kj

0.68

Activations Density 0.113%

phrases indicating communication or contact

instances of communication or messages from various sources

No Comments

No Known Activations