INDEX

Explanations

words related to making changes or modifications

oai_token-act-pair · gpt-3.5-turbo

terms related to modifications or changes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

enic

-0.80

 Clancy

-0.71

gil

-0.70

 Vide

-0.67

VP

-0.67

ILE

-0.66

 Chrys

-0.66

ardy

-0.65

jay

-0.65

ocry

-0.64

POSITIVE LOGITS

 adjustment

1.50

 adjustments

1.40

adjust

1.10

 adjust

1.10

 adjusting

1.03

 Adjust

0.98

 thresholds

0.95

aution

0.95

 adjusts

0.93

Adjust

0.89

Activations Density 0.005%

words related to making changes or modifications

terms related to modifications or changes

No Comments

No Known Activations