INDEX

Explanations

terms related to health and safety

oai_token-act-pair · gpt-3.5-turbo

terms associated with health, safety, and community welfare

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

é¾į

-0.72

uden

-0.66

upper

-0.62

è£ħ

-0.61

 Pillar

-0.61

>>\

-0.60

emen

-0.58

election

-0.58

 Passage

-0.57

POSITIVE LOGITS

wise

0.79

etc

0.75

 permitting

0.65

ardless

0.63

aila

0.63

 versa

0.62

etheless

0.61

 biologists

0.58

 notwithstanding

0.58

 allocations

0.57

Activations Density 0.671%

terms related to health and safety

terms associated with health, safety, and community welfare

No Comments

No Known Activations