INDEX

Explanations

phrases referring to choices or options

oai_token-act-pair · gpt-3.5-turbo

the word "which" as it relates to clauses providing additional information

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

Behind

-0.72

athi

-0.70

Bas

-0.68

Ott

-0.64

grim

-0.62

STE

-0.62

 Showdown

-0.61

Buc

-0.60

 Passage

-0.60

UG

-0.60

POSITIVE LOGITS

 resulted

0.86

soever

0.85

allows

0.80

 admittedly

0.80

 brings

0.79

 incidentally

0.79

 includes

0.79

 culminated

0.78

milo

0.77

 consists

0.77

Activations Density 0.134%

phrases referring to choices or options

the word "which" as it relates to clauses providing additional information

No Comments

No Known Activations