INDEX

Explanations

references to photographs or images

oai_token-act-pair · gpt-3.5-turbo

instances of the word "photos" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ŃĶ

-0.66

 foreseeable

-0.63

 Rutherford

-0.62

mental

-0.62

bilt

-0.61

CFR

-0.60

osi

-0.60

owder

-0.60

edIn

-0.59

 Proposition

-0.58

POSITIVE LOGITS

ynthesis

1.72

ynt

1.54

ensitive

1.35

que

1.19

 depicting

1.06

ensitivity

1.05

hots

1.02

avers

0.99

 depict

0.94

creen

0.92

Activations Density 0.063%

references to photographs or images

instances of the word "photos" and related terms

No Comments

No Known Activations