INDEX

Explanations

numeric sequences such as dates, numbers, and codes

oai_token-act-pair · gpt-3.5-turbo

references to numerical indicators or counts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

gerald

-0.73

manship

-0.61

 "$:/

-0.61

awaru

-0.59

uine

-0.59

utan

-0.58

rolet

-0.58

ians

-0.58

urst

-0.56

ured

-0.56

POSITIVE LOGITS

nd

2.06

ND

1.17

1.04

1.03

1.02

 thirds

0.97

 externalToEVAOnly

0.96

0.95

0.92

0.91

Activations Density 0.135%

numeric sequences such as dates, numbers, and codes

references to numerical indicators or counts

No Comments

No Known Activations