INDEX

Explanations

different kinds or types of objects or concepts

oai_token-act-pair · gpt-3.5-turbo

references to different categories or classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Tycoon

-0.76

NING

-0.69

WN

-0.68

âĸ¬

-0.64

IRO

-0.64

 Thumbnails

-0.62

heid

-0.61

UTERS

-0.60

ned

-0.60

ITED

-0.59

POSITIVE LOGITS

etting

1.43

etter

1.18

pace

1.08

paces

1.08

uit

0.99

hell

0.94

uits

0.93

hips

0.91

afe

0.85

hots

0.85

Activations Density 0.056%

different kinds or types of objects or concepts

references to different categories or classifications

No Comments

No Known Activations