INDEX

Explanations

phrases related to ongoing events or continual actions

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

ItemTracker

-0.74

emia

-0.73

Obj

-0.69

ribute

-0.66

essee

-0.65

 Beir

-0.64

raint

-0.64

ractor

-0.63

Reviewed

-0.62

ouri

-0.62

POSITIVE LOGITS

theless

1.17

green

1.06

lasting

1.02

THING

0.93

thing

0.93

 dreamed

0.85

more

0.84

bilt

0.75

 wondered

0.73

Quest

0.73

Activations Density 0.260%

phrases related to ongoing events or continual actions

No Comments

No Known Activations