INDEX

Explanations

phrases related to specific names or titles

oai_token-act-pair · gpt-3.5-turbo

phrases that contain comma-separated lists and descriptors in a general context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

icultural

-0.76

hesda

-0.70

 disqualified

-0.66

 takeaway

-0.66

 plunge

-0.65

 barriers

-0.65

 bleeding

-0.64

 immedi

-0.62

 attrition

-0.62

iaries

-0.61

POSITIVE LOGITS

 coined

0.83

aka

0.81

alias

0.77

Boh

0.76

pron

0.75

 Definitions

0.73

lain

0.71

SourceFile

0.70

 abbre

0.70

 abbrevi

0.70

Activations Density 0.298%

phrases related to specific names or titles

phrases that contain comma-separated lists and descriptors in a general context

No Comments

No Known Activations