INDEX

Explanations

references to hidden or confidential information

oai_token-act-pair · gpt-3.5-turbo

references to secrets

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

annis

-0.82

odcast

-0.82

©¶æ

-0.81

gaard

-0.78

lihood

-0.77

thood

-0.75

FG

-0.75

chairs

-0.74

adjusted

-0.74

ulf

-0.74

POSITIVE LOGITS

ariat

1.02

arial

0.97

 ingredient

0.91

 handshake

0.86

rets

0.85

 secret

0.82

 hatch

0.80

 underground

0.78

 tunnel

0.78

 stash

0.77

Activations Density 0.017%

references to hidden or confidential information

references to secrets

No Comments

No Known Activations