INDEX

Explanations

the word "refused" followed by either "to" or "to" with another verb

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

è¦ļéĨĴ

-0.80

gone

-0.74

ulative

-0.73

isf

-0.73

ipel

-0.71

########

-0.71

atures

-0.71

ammy

-0.70

psc

-0.70

utenberg

-0.69

POSITIVE LOGITS

 extradition

0.93

 outright

0.87

 permission

0.81

 admission

0.79

 requests

0.77

 vehemently

0.75

adm

0.73

 altogether

0.73

 invitations

0.71

 answering

0.69

Activations Density 0.525%

the word "refused" followed by either "to" or "to" with another verb

No Comments

No Known Activations