INDEX

Explanations

specific time references

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

phrases that indicate specific times or temporal references

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 respectively

-0.77

soType

-0.67

encers

-0.63

 integrates

-0.62

 surpassed

-0.60

Reports

-0.59

å§«

-0.58

Released

-0.57

selves

-0.57

 Likewise

-0.56

POSITIVE LOGITS

 first

0.97

 least

0.97

 length

0.94

 some

0.93

roph

0.90

 times

0.87

first

0.87

one

0.86

 halftime

0.83

onement

0.82

Activations Density 0.104%

specific time references

phrases that indicate specific times or temporal references

No Known Activations