INDEX

Explanations

technical terms and codes that likely pertain to a specific field or industry

oai_token-act-pair · gpt-3.5-turbo

references to procedural or formal guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ĸļ

-0.75

 hashing

-0.71

 diving

-0.69

 scrap

-0.68

 nightly

-0.68

 stars

-0.66

 directly

-0.64

 hiking

-0.64

 ambassadors

-0.63

 consulate

-0.62

POSITIVE LOGITS

Tags

1.04

Added

1.02

Attach

1.01

Write

1.01

________________________________

0.98

Reader

0.97

Disclaimer

0.97

________________________________________________________________

0.97

Copyright

0.97

————————————————

0.96

Activations Density 0.507%

technical terms and codes that likely pertain to a specific field or industry

references to procedural or formal guidelines

No Comments

No Known Activations