INDEX

Explanations

exaggerated claims or estimations

oai_token-act-pair · gpt-3.5-turbo

terms related to estimation and evaluation, specifically underestimating and overestimating

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ĪĴ

-0.79

iety

-0.68

 Skydragon

-0.67

iston

-0.65

aine

-0.65

ŃĶ

-0.63

alion

-0.62

alde

-0.62

aining

-0.61

 thru

-0.60

POSITIVE LOGITS

imates

1.09

imate

1.02

imating

0.98

imated

0.95

 underest

0.94

 overest

0.94

imation

0.93

 underestimate

0.83

 msec

0.82

 exagger

0.81

Activations Density 0.091%

exaggerated claims or estimations

terms related to estimation and evaluation, specifically underestimating and overestimating

No Comments

No Known Activations