INDEX

Explanations

dates in the format of "Month Day Year" where the day and year are activated

oai_token-act-pair · gpt-3.5-turbo

dates and time-related information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

chy

-0.66

ufact

-0.63

leneck

-0.60

orable

-0.60

ioch

-0.58

 Arche

-0.57

ãĤ¼

-0.57

eton

-0.57

cules

-0.55

iaries

-0.54

POSITIVE LOGITS

th

1.12

eteenth

0.89

rd

0.80

roth

0.77

0.73

ember

0.73

0.72

venth

0.72

0.71

Activations Density 0.109%

dates in the format of "Month Day Year" where the day and year are activated

dates and time-related information

No Comments

No Known Activations