INDEX

Explanations

instances where something is being evaluated or judged

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate societal judgments or evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Ring

-0.72

Jet

-0.71

inas

-0.70

 Lawn

-0.67

 Planes

-0.67

 Observatory

-0.64

 Rhythm

-0.62

 interpreter

-0.60

 Alley

-0.60

 vocals

-0.60

POSITIVE LOGITS

 considered

0.99

aimon

0.88

 consider

0.87

ilitarian

0.86

phas

0.86

 favorably

0.80

agine

0.80

 considers

0.77

onyms

0.77

erala

0.75

Activations Density 0.014%

instances where something is being evaluated or judged

phrases that indicate societal judgments or evaluations

No Comments

No Known Activations