INDEX

Explanations

specific mentions of the word "audit"

oai_token-act-pair · gpt-3.5-turbo

terms related to audits and accounting

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

theless

-0.80

ulkan

-0.77

————————————————

-0.72

otine

-0.71

axe

-0.71

ktop

-0.65

onz

-0.64

fires

-0.63

joy

-0.63

pite

-0.63

POSITIVE LOGITS

 audit

1.07

yssey

0.88

 audits

0.87

onomy

0.80

 auditor

0.79

 reinvest

0.77

orate

0.75

idated

0.73

 appraisal

0.73

 probing

0.72

Activations Density 0.018%

specific mentions of the word "audit"

terms related to audits and accounting

No Comments

No Known Activations