INDEX

Explanations

phrases related to specific parts or sections of something

oai_token-act-pair · gpt-3.5-turbo

references to roles or positions within a group or context

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

terday

-0.85

aukee

-0.80

ICO

-0.79

ãĤ´ãĥ³

-0.75

lishes

-0.72

jriwal

-0.69

 Rampage

-0.68

ccording

-0.65

 Despair

-0.65

hammad

-0.65

POSITIVE LOGITS

 eyed

0.74

osphere

0.73

lights

0.67

ales

0.63

 closest

0.62

 furthe

0.59

als

0.57

 nearest

0.57

guy

0.56

iverse

0.56

Activations Density 0.182%

phrases related to specific parts or sections of something

references to roles or positions within a group or context

No Comments

No Known Activations