INDEX

Explanations

technical terms or phrases related to sophistication, quality, or advanced classification

oai_token-act-pair · gpt-3.5-turbo

terms related to classification or grading systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ynt

-0.87

oleon

-0.84

ingle

-0.81

pid

-0.79

Stone

-0.78

pload

-0.76

Wil

-0.76

SN

-0.75

idental

-0.75

ypes

-0.74

POSITIVE LOGITS

 immersion

0.74

 alternative

0.72

 exotic

0.71

 garbage

0.70

 enterprise

0.69

 entertainment

0.69

 behavior

0.68

 insanity

0.68

 coercion

0.67

 extreme

0.67

Activations Density 0.221%

technical terms or phrases related to sophistication, quality, or advanced classification

terms related to classification or grading systems

No Comments

No Known Activations