INDEX

Explanations

references to elite groups or individuals

oai_token-act-pair · gpt-3.5-turbo

the word "elite" in various contexts related to status or group 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

SourceFile

-0.75

asca

-0.72

DAY

-0.69

OUT

-0.69

asta

-0.68

aul

-0.68

alam

-0.67

externalActionCode

-0.65

pty

-0.65

 Parables

-0.65

POSITIVE LOGITS

 elite

0.84

 Institution

0.78

 athlete

0.77

 athletes

0.75

 credentials

0.75

 mosqu

0.73

holm

0.73

 acad

0.71

 academy

0.70

 classes

0.68

Activations Density 0.024%

references to elite groups or individuals

the word "elite" in various contexts related to status or group classifications

No Comments

No Known Activations