INDEX

Explanations

phrases related to individuals with medical or academic titles

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

 chorus

-0.67

CHAT

-0.66

 quart

-0.65

 Territories

-0.64

 curfew

-0.63

 spirited

-0.63

 routed

-0.61

 compens

-0.60

 rout

-0.60

 coupling

-0.60

POSITIVE LOGITS

inks

1.11

unks

1.07

umin

1.05

inker

1.05

illing

1.04

inking

1.02

astically

1.00

ifts

0.98

ifting

0.98

herty

0.97

Activations Density 0.255%

phrases related to individuals with medical or academic titles

No Comments

No Known Activations