INDEX

Explanations

references to "Big" followed by numbers, indicating a focus on sports conference names

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

istry

-0.84

idency

-0.78

 confir

-0.77

theless

-0.73

fully

-0.69

yrim

-0.68

 muster

-0.67

ILA

-0.67

cia

-0.67

lessness

-0.66

POSITIVE LOGITS

gest

1.46

ger

1.24

gie

1.05

 Brother

0.99

gins

0.93

gers

0.90

wig

0.89

 Daddy

0.87

 Bang

0.86

glers

0.85

Activations Density 0.018%

references to "Big" followed by numbers, indicating a focus on sports conference names

No Comments

No Known Activations