INDEX

Explanations

phrases indicating depth or analysis

oai_token-act-pair · gpt-3.5-turbo

references to depth or profoundness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Prosecut

-0.66

bill

-0.64

obe

-0.64

kill

-0.62

 Provided

-0.62

 Elect

-0.61

von

-0.59

 Volunteers

-0.58

cert

-0.58

cop

-0.58

POSITIVE LOGITS

 deeper

3.51

 deep

2.04

 deepest

1.95

 closer

1.71

 deepen

1.68

 richer

1.66

 deepening

1.66

 wider

1.65

 darker

1.57

eeper

1.55

Activations Density 0.010%

phrases indicating depth or analysis

references to depth or profoundness

No Comments

No Known Activations