INDEX

Explanations

prefixes such as "pre-" and words associated with importance or precedence

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "pre-" in words

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

ItemThumbnailImage

-0.68

ById

-0.65

ãĥīãĥ©ãĤ´ãĥ³

-0.64

Lago

-0.64

ĸļ

-0.61

 independents

-0.61

oub

-0.60

 needle

-0.60

=-=-=-=-

-0.60

ioxide

-0.59

POSITIVE LOGITS

atural

0.85

emption

0.81

atives

0.75

existing

0.74

itional

0.73

ursed

0.73

ussions

0.71

aturally

0.71

utions

0.69

ition

0.69

Activations Density 0.046%

prefixes such as "pre-" and words associated with importance or precedence

occurrences of the prefix "pre-" in words

No Comments

No Known Activations