INDEX

Explanations

invitations or prompts for engagement

oai_token-act-pair · gpt-3.5-turbo

calls to action and requests for engagement

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

isp

-0.63

 glyph

-0.63

meta

-0.61

adra

-0.61

 footh

-0.60

CTR

-0.60

£ı

-0.59

eco

-0.59

native

-0.59

spir

-0.59

POSITIVE LOGITS

?????-?????-

0.90

Answer

0.84

 Nope

0.77

.?

0.76

--------------------------------------------------------

0.72

 Answer

0.71

 Virgin

0.69

?????-

0.69

 kidding

0.69

emate

0.68

Activations Density 0.238%

invitations or prompts for engagement

calls to action and requests for engagement

No Comments

No Known Activations