INDEX

Explanations

references to people involved in various incidents or 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

ooks

-0.74

akings

-0.72

eeds

-0.71

urches

-0.70

poons

-0.68

ernels

-0.68

brids

-0.68

 Cups

-0.68

 gears

-0.66

 scripts

-0.66

POSITIVE LOGITS

who

1.17

 whom

1.11

 named

1.02

 whose

1.00

who

0.88

 classmate

0.86

friend

0.85

 colleague

0.79

 friend

0.77

 acquaintance

0.77

Activations Density 0.119%

references to people involved in various incidents or contexts

No Comments

No Known Activations