INDEX

Explanations

technical terms and programming-related instructions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the state or condition of parameters or variables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

isine

-0.86

Â£

-0.81

Pros

-0.79

edin

-0.75

rave

-0.73

 Breach

-0.69

staking

-0.69

ventures

-0.68

Vers

-0.68

Purchase

-0.66

POSITIVE LOGITS

 initialized

1.42

 parsed

1.33

 evaluated

1.26

 invoked

1.24

 mapped

1.20

 computed

1.20

 reused

1.19

 interpreted

1.18

 defined

1.17

 executed

1.16

Activations Density 0.094%

technical terms and programming-related instructions

phrases indicating the state or condition of parameters or variables

No Comments

No Known Activations