INDEX

Explanations

the name "Waters" with different levels of emphasis, likely indicating relevance or importance in the context of the text

oai_token-act-pair · gpt-3.5-turbo

mentions of the name 'Waters'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

oard

-0.70

ERAL

-0.69

bable

-0.68

erous

-0.66

 conviction

-0.62

obs

-0.62

OTS

-0.62

 consideration

-0.62

 conceal

-0.61

 centralized

-0.61

POSITIVE LOGITS

 Waters

1.19

waters

0.96

melon

0.94

bats

0.92

hed

0.85

hip

0.85

combe

0.83

boro

0.82

pin

0.82

geist

0.82

Activations Density 0.006%

the name "Waters" with different levels of emphasis, likely indicating relevance or importance in the context of the text

mentions of the name 'Waters'

No Comments

No Known Activations