INDEX

Explanations

dates in the format of four digits representing years with high activation values

oai_token-act-pair · gpt-3.5-turbo

year references in historical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

amins

-0.67

agy

-0.61

 Gods

-0.61

 genus

-0.59

elong

-0.59

worldly

-0.58

ãģį

-0.57

 unpredictable

-0.57

icons

-0.57

atio

-0.57

POSITIVE LOGITS

 onwards

0.84

 partName

0.82

0.78

 sidx

0.68

 Accord

0.66

 Kass

0.64

 when

0.63

 onward

0.62

assetsadobe

0.61

EV

0.60

Activations Density 0.085%

dates in the format of four digits representing years with high activation values

year references in historical contexts

No Comments

No Known Activations