INDEX

Explanations

technical terms or jargon related to software development

oai_token-act-pair · gpt-3.5-turbo

references to requests for information or contributions from others

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 contrace

-0.70

 fortun

-0.62

 Tsukuyomi

-0.61

llah

-0.61

elfare

-0.61

 livest

-0.59

 discipl

-0.59

ptives

-0.58

xon

-0.57

aos

-0.56

POSITIVE LOGITS

 overlooked

0.94

 missed

0.92

 missing

0.91

 noteworthy

0.91

 favorites

0.85

 standout

0.79

 omitted

0.77

 interesting

0.77

missing

0.77

 worth

0.76

Activations Density 0.526%

technical terms or jargon related to software development

references to requests for information or contributions from others

No Comments

No Known Activations