INDEX

Explanations

pronouns, particularly those associated with individuals, indicating their actions or states

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

 Marriage

-0.69

 Fashion

-0.63

duc

-0.63

 Domestic

-0.63

 Dominion

-0.62

 apex

-0.60

 Sussex

-0.58

ogue

-0.58

piring

-0.57

Gad

-0.57

POSITIVE LOGITS

'd

1.14

 personally

1.02

 encount

1.02

 consulted

0.98

 regretted

0.97

 awoke

0.96

'll

0.93

 hoped

0.89

've

0.85

 heard

0.85

Activations Density 0.097%

pronouns, particularly those associated with individuals, indicating their actions or states

No Comments

No Known Activations