INDEX

Explanations

references to political figures and their interactions in a diplomatic context

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

Ö¼

-0.66

 phased

-0.61

 ILCS

-0.61

 Dangerous

-0.60

ucket

-0.58

Wr

-0.57

]=

-0.56

 proportion

-0.55

 commodity

-0.55

 subscrib

-0.54

POSITIVE LOGITS

 representatives

0.90

armac

0.78

ilaterally

0.76

 regarding

0.75

strate

0.75

 backstage

0.74

 counterparts

0.73

 Listen

0.73

 discussing

0.72

reet

0.72

Activations Density 0.242%

references to political figures and their interactions in a diplomatic context

No Comments

No Known Activations