INDEX

Explanations

references to the word "tin" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

enance

-0.95

 Emanuel

-0.67

 Gamergate

-0.65

dL

-0.65

 Vector

-0.62

 Dayton

-0.62

ean

-0.61

 Lynn

-0.61

 SYSTEM

-0.61

 Wiley

-0.61

POSITIVE LOGITS

kered

1.32

icum

1.16

ned

1.16

oco

1.15

sel

1.08

iqu

0.98

 oxide

0.98

ocular

0.95

omial

0.94

ny

0.94

Activations Density 0.004%

references to the word "tin" in various contexts

No Comments

No Known Activations