INDEX

Explanations

phrases related to trends and renditions

oai_token-act-pair · gpt-3.5-turbo

terms associated with rendering or visual representation in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

OPS

-0.69

 Kaufman

-0.66

aiman

-0.65

©¶æ

-0.64

avez

-0.64

keley

-0.64

«ĺ

-0.63

 cumbers

-0.63

Magikarp

-0.62

 Lovecraft

-0.61

POSITIVE LOGITS

ezvous

1.54

rend

1.22

erer

1.22

erers

1.05

eering

1.03

ing

0.96

ered

0.94

lyn

0.87

ers

0.84

ings

0.84

Activations Density 0.009%

phrases related to trends and renditions

terms associated with rendering or visual representation in various contexts

No Comments

No Known Activations