INDEX

Explanations

phrases related to secrets, hidden information, or concealed objects

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "hidden" or "hiddenness."

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

-0.86

nia

-0.84

xual

-0.81

Pwr

-0.79

ctive

-0.77

SHIP

-0.76

ocations

-0.74

udeb

-0.73

ufact

-0.73

urses

-0.73

POSITIVE LOGITS

 hidden

0.91

 Hidden

0.84

 behind

0.83

 recess

0.82

door

0.78

 stash

0.77

 hatch

0.77

gem

0.76

 tunnel

0.76

 treasure

0.75

Activations Density 0.014%

phrases related to secrets, hidden information, or concealed objects

references to the concept of "hidden" or "hiddenness."

No Comments

No Known Activations