INDEX

Explanations

individuals by their names

oai_token-act-pair · gpt-3.5-turbo

instances of the word "who" referring to individuals in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ulp

-0.68

Nec

-0.62

utation

-0.62

DDR

-0.60

 Affordable

-0.60

 Bound

-0.59

AGE

-0.59

ķ

-0.58

Georg

-0.58

reach

-0.58

POSITIVE LOGITS

 oversaw

0.90

 preceded

0.89

anwhile

0.88

 resided

0.88

 accompanies

0.88

 owns

0.87

soever

0.87

 oversees

0.85

 famously

0.85

 incidentally

0.84

Activations Density 0.093%

individuals by their names

instances of the word "who" referring to individuals in various contexts

No Comments

No Known Activations