INDEX

Explanations

the word "that" 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

 Cosponsors

-0.73

ãĥ³ãĤ¸

-0.73

english

-0.71

icons

-0.69

ographies

-0.69

 Shares

-0.65

oway

-0.65

obal

-0.65

LLOW

-0.65

çļ

-0.64

POSITIVE LOGITS

 fateful

1.07

 pesky

0.93

 happens

0.88

 debacle

0.87

 happened

0.87

 bombshell

0.83

 initial

0.83

 fiasco

0.81

 incident

0.80

 period

0.76

Activations Density 0.022%

the word "that" in various contexts

No Comments

No Known Activations