INDEX

Explanations

specific phrases denoting a continuation of a sequence or list

oai_token-act-pair · gpt-3.5-turbo

the word "following" in various contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

rimination

-0.78

cci

-0.77

aucas

-0.72

iman

-0.71

adle

-0.71

elli

-0.70

etime

-0.70

otor

-0.69

obyl

-0.69

eri

-0.69

POSITIVE LOGITS

 follows

0.80

 week

0.69

 weeks

0.69

 newsp

0.68

 following

0.67

 follow

0.64

 sessions

0.63

ĸļ

0.63

combe

0.62

 sections

0.62

Activations Density 0.023%

specific phrases denoting a continuation of a sequence or list

the word "following" in various contexts

No Known Activations