INDEX

Explanations

numerical targets or figures mentioned in relation to goals or limits

oai_token-act-pair · gpt-3.5-turbo

mentions of targets and limits, particularly related to emissions or renewable energy

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

 istg

-0.66

Interstitial

-0.65

 sidel

-0.61

 VIDEOS

-0.61

successfully

-0.60

 Everywhere

-0.60

Struct

-0.59

aband

-0.59

actionDate

-0.58

politics

-0.58

POSITIVE LOGITS

 estimate

0.95

onement

0.86

 number

0.85

 amount

0.85

 allowance

0.85

 (Â£

0.83

 expires

0.82

 allowable

0.80

 estimates

0.79

emaker

0.77

Activations Density 0.200%

numerical targets or figures mentioned in relation to goals or limits

mentions of targets and limits, particularly related to emissions or renewable energy

No Comments

No Known Activations