INDEX

Explanations

words indicating interesting or noteworthy information

oai_token-act-pair · gpt-3.5-turbo

phrases that start with "Interestingly" or "Ironically."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 newsletters

-0.69

 throats

-0.66

"},"

-0.65

 throat

-0.64

 mouths

-0.64

 fronts

-0.61

utm

-0.60

 meats

-0.58

 jaws

-0.57

 speech

-0.57

POSITIVE LOGITS

Interestingly

0.78

 interestingly

0.70

 paradox

0.70

 Interestingly

0.70

ety

0.67

Ironically

0.65

 Ironically

0.65

SPONSORED

0.65

bley

0.63

Corp

0.63

Activations Density 0.020%

words indicating interesting or noteworthy information

phrases that start with "Interestingly" or "Ironically."

No Comments

No Known Activations