INDEX

Explanations

references to quantities and comparisons 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

ousse

-0.66

fed

-0.60

Staff

-0.60

tops

-0.58

item

-0.58

urities

-0.57

ensitive

-0.55

pd

-0.55

Desktop

-0.54

REP

-0.54

POSITIVE LOGITS

 route

0.93

 furthe

0.80

bye

0.73

 Definitive

0.70

 mile

0.69

 downhill

0.69

 distance

0.69

 step

0.66

 lengths

0.65

ither

0.64

Activations Density 0.050%

references to quantities and comparisons in various contexts

No Comments

No Known Activations