INDEX

Explanations

adjectives and phrases expressing surprise or interest

oai_token-act-pair · gpt-3.5-turbo

phrases that express unusual or surprising observations

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

 throats

-0.68

ujah

-0.65

ona

-0.63

 throat

-0.60

tering

-0.60

eering

-0.59

shaw

-0.58

ailability

-0.58

taboola

-0.57

 refriger

-0.57

POSITIVE LOGITS

 juxtap

0.73

SPONSORED

0.71

iasco

0.70

 excerpts

0.69

 EVEN

0.69

 Marino

0.68

 Rasmussen

0.66

 remem

0.65

 interestingly

0.64

 contrasting

0.63

Activations Density 0.221%

adjectives and phrases expressing surprise or interest

phrases that express unusual or surprising observations

No Comments

No Known Activations