INDEX

Explanations

terms related to decreases or declines in metrics, particularly in economic 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

urally

-0.78

 extras

-0.66

 subtitles

-0.64

 quir

-0.64

rich

-0.62

 redacted

-0.62

Sources

-0.61

ACTED

-0.61

REDACTED

-0.61

killed

-0.60

POSITIVE LOGITS

lust

0.95

down

0.94

downs

0.93

 rate

0.92

over

0.88

backs

0.84

 precip

0.83

nings

0.82

back

0.79

ait

0.79

Activations Density 0.076%

terms related to decreases or declines in metrics, particularly in economic contexts

No Comments

No Known Activations