INDEX

Explanations

references to individuals or groups of people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tnc

-0.88

 srfAttach

-0.87

Accessory

-0.78

COMPLE

-0.72

NES

-0.70

uner

-0.69

efully

-0.68

iary

-0.64

Prov

-0.64

CCC

-0.64

POSITIVE LOGITS

 smugglers

1.04

who

1.00

folk

0.95

 underestimate

0.84

 perceive

0.83

 wanting

0.83

 clam

0.82

 else

0.81

 prefer

0.78

 cared

0.77

Activations Density 0.104%

references to individuals or groups of people

No Comments

No Known Activations