INDEX

Explanations

dates related to the year 2009

oai_token-act-pair · gpt-3.5-turbo

mentions of the year 2009

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

onite

-0.89

amins

-0.72

heid

-0.71

onal

-0.71

orable

-0.69

ealing

-0.66

aban

-0.64

kef

-0.62

alty

-0.62

esan

-0.62

POSITIVE LOGITS

å¹

0.80

-'

0.80

0.76

0.73

0.70

 onwards

0.68

0.67

0.66

master

0.65

0.64

Activations Density 0.014%

dates related to the year 2009

mentions of the year 2009

No Comments

No Known Activations