INDEX

Explanations

information related to different projects, often pertaining to photography, companies, and organizations

oai_token-act-pair · gpt-3.5-turbo

iterative phrases that emphasize personal introductions or descriptions

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

 wiser

-0.68

icals

-0.66

 phased

-0.63

 corresponding

-0.61

 guessing

-0.61

£ı

-0.61

 substitutes

-0.59

 commuting

-0.59

 Aless

-0.56

 equivalents

-0.55

POSITIVE LOGITS

Background

0.89

========

0.77

Kick

0.76

Hello

0.73

Object

0.72

Hi

0.71

atography

0.70

BACK

0.70

This

0.69

↵

0.68

Activations Density 0.093%

information related to different projects, often pertaining to photography, companies, and organizations

iterative phrases that emphasize personal introductions or descriptions

No Comments

No Known Activations