INDEX

Explanations

mentions of a specific subject throughout the document

oai_token-act-pair · gpt-3.5-turbo

references to topics and issues of discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

PB

-0.63

othing

-0.62

lett

-0.62

ADRA

-0.61

eria

-0.60

OUP

-0.59

dp

-0.59

aukee

-0.58

 Towers

-0.57

igi

-0.57

POSITIVE LOGITS

of

0.89

 thereof

0.79

 itself

0.75

atics

0.71

href

0.68

ophys

0.66

 posed

0.65

 raised

0.65

fulness

0.64

naires

0.64

Activations Density 0.130%

mentions of a specific subject throughout the document

references to topics and issues of discussion

No Comments

No Known Activations