INDEX

Explanations

instances where a location or entity serves a specific purpose or function

oai_token-act-pair · gpt-3.5-turbo

phrases that denote functionality or roles of subjects within a context

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

rants

-0.74

ulous

-0.71

stro

-0.69

 Rebell

-0.69

 Transactions

-0.61

Pic

-0.61

ties

-0.60

 Believe

-0.60

 Reasons

-0.60

Ç

-0.60

POSITIVE LOGITS

 liaison

0.90

 adjunct

0.88

 interim

0.86

 lookout

0.84

 fodder

0.81

 intermediary

0.81

 invaluable

0.77

 reminder

0.77

 antit

0.76

 ambassadors

0.75

Activations Density 0.094%

instances where a location or entity serves a specific purpose or function

phrases that denote functionality or roles of subjects within a context

No Comments

No Known Activations