INDEX

Explanations

actions that happen subsequently or eventually

oai_token-act-pair · gpt-3.5-turbo

adverbs indicating the timing or sequence of events

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

 VIDEOS

-0.81

ggles

-0.69

 Increases

-0.66

 sucks

-0.66

 weekday

-0.65

insula

-0.65

 circles

-0.63

 meshes

-0.63

Ago

-0.61

doms

-0.61

POSITIVE LOGITS

 able

0.89

 supposed

0.87

 considered

0.82

minent

0.82

done

0.79

 subjected

0.79

 ready

0.77

 banned

0.77

 aware

0.76

 subject

0.75

Activations Density 0.271%

actions that happen subsequently or eventually

adverbs indicating the timing or sequence of events

No Comments

No Known Activations