INDEX

Explanations

references to items that are not firmly attached or established

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being "loose" in various contexts

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

ãĤ´ãĥ³

-0.81

aurus

-0.78

HCR

-0.76

amus

-0.74

IENCE

-0.74

ophers

-0.73

issance

-0.72

illian

-0.71

igraph

-0.70

urai

-0.70

POSITIVE LOGITS

leaf

1.14

cloth

0.94

tail

0.86

tails

0.85

weed

0.84

luster

0.83

bound

0.82

fitting

0.82

ness

0.80

wd

0.79

Activations Density 0.016%

references to items that are not firmly attached or established

references to the concept of being "loose" in various contexts

No Comments

No Known Activations