INDEX

Explanations

images or photographs mentioned in a text

oai_token-act-pair · gpt-3.5-turbo

references to images or photos being shared or posted

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

merce

-0.90

schild

-0.90

ensable

-0.81

endo

-0.74

lear

-0.72

ENDED

-0.72

osponsors

-0.70

DEF

-0.70

EED

-0.69

gaard

-0.68

POSITIVE LOGITS

 depicting

1.01

 caption

0.88

 Snapchat

0.79

 photograph

0.76

 portraying

0.74

 photo

0.74

 screenshot

0.74

 showing

0.73

journal

0.72

img

0.72

Activations Density 0.053%

images or photographs mentioned in a text

references to images or photos being shared or posted

No Comments

No Known Activations