INDEX

Explanations

the name "Warren" 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

dayName

-0.84

netflix

-0.76

ãĤ´ãĥ³

-0.75

odic

-0.72

 conflic

-0.71

asus

-0.69

ãĤº

-0.67

alogue

-0.66

mble

-0.66

cific

-0.65

POSITIVE LOGITS

 Buffett

1.19

sburg

0.98

 Farrell

0.96

rade

0.89

 Warren

0.86

 Sapp

0.85

 Harding

0.81

ston

0.81

ers

0.80

 Davidson

0.79

Activations Density 0.005%

the name "Warren" in various contexts

No Comments

No Known Activations