INDEX

Explanations

numerical data and statistics in the context of specifications or ratings

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

¥ŀ

-0.66

rame

-0.65

 destro

-0.63

 crush

-0.61

 wider

-0.60

versions

-0.60

 widening

-0.60

 forcefully

-0.59

 broader

-0.59

 wiping

-0.58

POSITIVE LOGITS

 Types

0.79

0.78

0.77

 Months

0.77

Activations Density 0.049%

numerical data and statistics in the context of specifications or ratings

No Comments

No Known Activations