INDEX

Explanations

disruption-related terms and names of individuals or companies related to disruptive activities

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

 Juliet

-0.68

 Burr

-0.66

glers

-0.62

 hell

-0.62

 Cullen

-0.61

hetti

-0.61

 Rutherford

-0.61

 Deer

-0.61

Hod

-0.60

 Lans

-0.59

POSITIVE LOGITS

burse

1.19

cipl

1.19

cerning

1.16

rup

1.15

qus

1.15

ruption

1.14

interested

1.12

section

1.08

enfranch

1.07

array

1.07

Activations Density 0.364%

disruption-related terms and names of individuals or companies related to disruptive activities

No Comments

No Known Activations