INDEX

Explanations

words related to invisibility or being unseen

oai_token-act-pair · gpt-3.5-turbo

terms related to invisibility or being unnoticed

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

olitan

-0.83

âķĲâķĲ

-0.79

xit

-0.79

odcast

-0.78

ortment

-0.77

ership

-0.77

sych

-0.76

aeper

-0.76

anchester

-0.76

ourses

-0.73

POSITIVE LOGITS

 invisible

0.88

 invis

0.85

 unseen

0.77

 volcano

0.70

 Invisible

0.69

 enemy

0.69

 phantom

0.68

judicial

0.67

 silhou

0.65

worm

0.65

Activations Density 0.024%

words related to invisibility or being unseen

terms related to invisibility or being unnoticed

No Comments

No Known Activations