INDEX

Explanations

phrases related to discovering or uncovering information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the act of discovering or learning information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

oun

-0.75

Crystal

-0.69

ovich

-0.69

avorite

-0.66

ĸļ

-0.64

luster

-0.64

berus

-0.62

eries

-0.61

uga

-0.59

erate

-0.58

POSITIVE LOGITS

 about

1.02

why

0.98

how

0.97

WHY

0.87

 whats

0.86

 beforehand

0.85

 exactly

0.85

 what

0.83

 whether

0.82

 afterwards

0.82

Activations Density 0.022%

phrases related to discovering or uncovering information

phrases indicating the act of discovering or learning information

No Comments

No Known Activations