INDEX

Explanations

statements about potential future events and outcomes

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

eret

-0.90

 yesterday

-0.81

 earlier

-0.77

 beforehand

-0.73

 recent

-0.71

 lately

-0.66

 recently

-0.62

mist

-0.61

 haste

-0.61

camp

-0.60

POSITIVE LOGITS

 surpass

0.76

 someday

0.73

 reckoning

0.73

omed

0.72

ynasty

0.72

ependence

0.70

ortality

0.70

 extinct

0.68

ypes

0.67

alia

0.67

Activations Density 0.283%

statements about potential future events and outcomes

No Comments

No Known Activations