INDEX

Explanations

mentions of innovative ideas, products, or approaches

oai_token-act-pair · gpt-3.5-turbo

expressions related to innovation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

phe

-0.71

ada

-0.66

andra

-0.65

nery

-0.65

 Hume

-0.64

 Suff

-0.63

apa

-0.63

walking

-0.62

 Allaah

-0.61

 whence

-0.61

POSITIVE LOGITS

 minded

1.00

 entreprene

0.91

 innovations

0.85

 innovative

0.85

 innovation

0.85

minded

0.81

 innov

0.79

 collaborations

0.77

 breakthrough

0.75

ortun

0.72

Activations Density 0.015%

mentions of innovative ideas, products, or approaches

expressions related to innovation

No Comments

No Known Activations