INDEX

Explanations

instances where something falls short of a certain expectation or requirement

oai_token-act-pair · gpt-3.5-turbo

phrases indicating something is lacking or insufficient

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

otation

-0.72

cu

-0.72

wa

-0.67

iple

-0.65

enery

-0.63

dies

-0.62

iment

-0.61

PID

-0.61

imentary

-0.60

atu

-0.60

POSITIVE LOGITS

sighted

0.78

 sanity

0.76

 expectations

0.75

 conclusive

0.73

 acceptable

0.72

 satisfactory

0.71

 resembling

0.71

osponsors

0.70

 subsistence

0.69

ormal

0.69

Activations Density 0.150%

instances where something falls short of a certain expectation or requirement

phrases indicating something is lacking or insufficient

No Comments

No Known Activations