INDEX

Explanations

terms related to technical or official standards

oai_token-act-pair · gpt-3.5-turbo

references to safety or quality standards

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

seat

-0.77

ston

-0.77

iversary

-0.73

hs

-0.71

rontal

-0.71

resent

-0.71

ãĥİ

-0.68

 Rove

-0.68

ience

-0.67

gary

-0.67

POSITIVE LOGITS

 deviations

1.05

 standards

1.02

 deviation

0.95

 compliant

0.89

heet

0.85

 governing

0.85

 Enforcement

0.78

 Compliance

0.77

 bearer

0.76

 regulating

0.75

Activations Density 0.028%

terms related to technical or official standards

references to safety or quality standards

No Comments

No Known Activations