INDEX

Explanations

names of individuals, especially in the context of politics, cinema, and journalism

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 eternity

-0.59

 Palest

-0.59

$.

-0.58

 Kafka

-0.57

 Gods

-0.55

Els

-0.53

 Pieces

-0.53

 Pengu

-0.53

Quantity

-0.53

+.

-0.52

POSITIVE LOGITS

 reacted

1.24

 took

1.15

 went

1.13

has

1.13

 gave

1.10

 prevailed

1.09

 weighed

1.08

 appeared

1.08

 welcomed

1.07

 showed

1.07

Activations Density 2.961%

names of individuals, especially in the context of politics, cinema, and journalism

No Comments

No Known Activations