INDEX

Explanations

references to sources and citations within a text

oai_token-act-pair · gpt-3.5-turbo

citations or references in the text

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

 lumber

-0.72

 stone

-0.67

 bread

-0.66

 trib

-0.66

 stra

-0.66

 conduct

-0.66

 punch

-0.66

 leve

-0.65

 ramp

-0.65

 morning

-0.65

POSITIVE LOGITS

Sources

1.42

 Sources

1.20

ibliography

1.04

 References

0.98

 Locations

0.96

References

0.92

 Consult

0.90

achev

0.89

cyclopedia

0.86

 Figures

0.85

Activations Density 0.014%

references to sources and citations within a text

citations or references in the text

No Comments

No Known Activations