INDEX

Explanations

system commands and instructions

oai_token-act-pair · gpt-3.5-turbo

future tense verbs

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

engineering

-0.73

quished

-0.72

JD

-0.68

 mastering

-0.66

quickShipAvailable

-0.65

Isn

-0.65

traumatic

-0.62

 vocational

-0.62

soType

-0.62

 Sources

-0.62

POSITIVE LOGITS

 automatically

1.13

be

1.12

 gladly

1.10

 ALWAYS

1.07

 suffice

0.97

 usually

0.97

 appear

0.96

 always

0.96

 notify

0.95

NOT

0.94

Activations Density 0.164%

system commands and instructions

future tense verbs

No Comments

No Known Activations