INDEX

Explanations

verbs related to devaluing or diminishing something

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

-0.65

andum

-0.62

tein

-0.59

Ã«

-0.56

 avail

-0.56

 negatively

-0.55

¶ħ

-0.55

-0.54

Ta

-0.52

 EVENTS

-0.52

POSITIVE LOGITS

graded

1.43

grading

1.29

LOAD

1.27

pour

1.16

stairs

1.16

grades

1.14

hill

1.04

loaded

1.01

sized

0.99

played

0.95

Activations Density 0.034%

verbs related to devaluing or diminishing something

No Comments

No Known Activations