INDEX

Explanations

terms related to actions or attributes showing intention, expertise, risks, and observations

oai_token-act-pair · gpt-3.5-turbo

words and phrases that indicate strong physical actions or emotions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Mehran

-0.71

estate

-0.65

']

-0.63

"]=>

-0.62

 Recomm

-0.56

Reviewer

-0.56

]'

-0.55

 Kahn

-0.55

...]

-0.55

"]

-0.55

POSITIVE LOGITS

 accents

0.62

 scissors

0.61

 caveats

0.57

 lenses

0.57

 flowing

0.57

 hindsight

0.57

 flourish

0.56

acity

0.55

 linem

0.53

 gloves

0.53

Activations Density 1.094%

terms related to actions or attributes showing intention, expertise, risks, and observations

words and phrases that indicate strong physical actions or emotions

No Comments

No Known Activations