INDEX

Explanations

terms related to criminal incidents involving police reports

oai_token-act-pair · gpt-3.5-turbo

negations or the word "no" 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

dor

-0.68

romy

-0.66

iership

-0.63

ãĤ¼ãĤ¦ãĤ¹

-0.62

ean

-0.62

rex

-0.62

dry

-0.62

redit

-0.60

istant

-0.60

tic

-0.60

POSITIVE LOGITS

xious

1.07

 longer

0.98

 indication

0.92

 doubt

0.91

terday

0.86

 matter

0.85

 clue

0.84

 conceivable

0.82

 mention

0.81

 sooner

0.79

Activations Density 0.095%

terms related to criminal incidents involving police reports

negations or the word "no" in various contexts

No Comments

No Known Activations