INDEX

Explanations

file related keywords and instructions

oai_token-act-pair · gpt-3.5-turbo

references to output and input mechanisms in a technical context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ECK

-0.77

aps

-0.70

anova

-0.67

creen

-0.66

acebook

-0.65

osuke

-0.65

 Carolyn

-0.65

 Antiqu

-0.65

illac

-0.64

aldi

-0.63

POSITIVE LOGITS

ting

1.20

ted

1.18

 outputs

0.82

Stream

0.78

 output

0.76

 streams

0.75

 Output

0.74

output

0.72

rious

0.71

 impedance

0.70

Activations Density 0.043%

file related keywords and instructions

references to output and input mechanisms in a technical context

No Comments

No Known Activations