INDEX

Explanations

phrases related to stopping or halting an action

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate limits or stopping points

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

properties

-0.81

soDeliveryDate

-0.80

nikov

-0.69

KNOWN

-0.69

eton

-0.65

olars

-0.63

MAC

-0.62

çļ

-0.62

ppa

-0.62

nai

-0.61

POSITIVE LOGITS

 bothering

0.72

ãĤ®

0.69

steen

0.68

 bleeding

0.66

 indefinitely

0.63

assing

0.63

 boredom

0.63

0.62

 interfering

0.62

 whining

0.61

Activations Density 0.177%

phrases related to stopping or halting an action

phrases that indicate limits or stopping points

No Comments

No Known Activations