INDEX

Explanations

scientific terms

oai_token-act-pair · gpt-3.5-turbo

references to "Scientific" and its variations, indicating a focus on scientific content or nomenclature

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

drops

-0.75

onies

-0.71

izons

-0.70

oning

-0.70

GG

-0.67

Torrent

-0.67

othy

-0.66

 evict

-0.65

OE

-0.65

ILA

-0.64

POSITIVE LOGITS

 Scientific

1.27

ivity

0.91

orie

0.87

 Instruments

0.86

natureconservancy

0.85

 Publications

0.84

aceutical

0.81

ally

0.81

 Fiction

0.80

scientific

0.79

Activations Density 0.007%

scientific terms

references to "Scientific" and its variations, indicating a focus on scientific content or nomenclature

No Comments

No Known Activations