INDEX

Explanations

phrases or words related to showing or listing content, information, or items

oai_token-act-pair · gpt-3.5-turbo

references to visual representations or images

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ateurs

-0.76

ingham

-0.74

cession

-0.64

riots

-0.61

rell

-0.59

omers

-0.59

reciation

-0.58

ateur

-0.58

assing

-0.58

 Clicker

-0.58

POSITIVE LOGITS

 herein

1.08

 therein

0.94

 above

0.94

 below

0.93

 separately

0.92

 prominently

0.91

 here

0.87

 hereafter

0.80

 elsewhere

0.79

 onstage

0.77

Activations Density 0.164%

phrases or words related to showing or listing content, information, or items

references to visual representations or images

No Comments

No Known Activations