INDEX

Explanations

contexts involving the concept of being "unsealed" or "revealed" in various scenarios

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

TON

-0.74

ioch

-0.69

DS

-0.67

IAN

-0.66

 stocks

-0.65

ornia

-0.65

ivities

-0.62

iants

-0.62

SON

-0.61

 à¨

-0.60

POSITIVE LOGITS

aled

1.05

xy

1.00

wed

0.98

ated

0.93

gments

0.93

ller

0.92

eded

0.92

aling

0.91

gment

0.90

emed

0.89

Activations Density 0.004%

contexts involving the concept of being "unsealed" or "revealed" in various scenarios

No Comments

No Known Activations