INDEX

Explanations

numbers and data formats

oai_token-act-pair · gpt-3.5-turbo

numerical identifiers and measurements

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

 tatt

-0.72

 suspic

-0.71

icularly

-0.65

ancial

-0.64

 horm

-0.62

igi

-0.62

 sugg

-0.61

 condu

-0.60

 perspect

-0.60

 arrang

-0.59

POSITIVE LOGITS

verages

0.75

.............

0.74

raft

0.68

egal

0.65

Legend

0.65

 Â£

0.65

Avg

0.65

0.64

OTAL

0.63

Ĥİ

0.63

Activations Density 0.148%

numbers and data formats

numerical identifiers and measurements

No Comments

No Known Activations