INDEX

Explanations

statistical comparisons highlighting extremes 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

unes

-0.84

ĨĴ

-0.74

ribut

-0.73

ertain

-0.70

must

-0.68

 Journals

-0.65

POST

-0.65

subject

-0.64

ripp

-0.64

hemer

-0.64

POSITIVE LOGITS

 overall

0.92

 reputation

0.86

 beard

0.81

 tendency

0.81

 penchant

0.79

 versatility

0.79

 propensity

0.78

 clout

0.78

 longevity

0.77

 firepower

0.76

Activations Density 0.049%

statistical comparisons highlighting extremes in various contexts

No Comments

No Known Activations