INDEX

Explanations

instances of phrases that indicate a high degree of intensity or extremity

oai_token-act-pair · gpt-3.5-turbo

adverbs that convey intensity or emphasis

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Pope

-0.71

 partName

-0.67

 Pump

-0.63

OPS

-0.63

¶æ

-0.60

 è£ıè

-0.60

ħĭ

-0.59

BTC

-0.59

 citizenship

-0.58

 slideshow

-0.57

POSITIVE LOGITS

enos

0.78

ymm

0.77

ense

0.76

azon

0.74

ocating

0.71

argo

0.69

ranged

0.66

amus

0.66

imb

0.66

amazon

0.65

Activations Density 0.127%

instances of phrases that indicate a high degree of intensity or extremity

adverbs that convey intensity or emphasis

No Comments

No Known Activations