INDEX

Explanations

Neuron 4 does not appear to be looking for anything in particular, as it has not been activated by any sections of the given text

oai_token-act-pair · gpt-4-turbo

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

SHIP

-0.83

loe

-0.80

schild

-0.79

ablishment

-0.73

hower

-0.72

riers

-0.70

rance

-0.70

aternity

-0.70

rill

-0.68

eryl

-0.65

POSITIVE LOGITS

associated

0.72

quished

0.70

 ILCS

0.68

 Thrones

0.65

uned

0.64

mediated

0.64

 Frozen

0.62

Scroll

0.61

SPONSORED

0.61

Beg

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.