INDEX

Explanations

words related to rejection or refusal

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reject" and its derivatives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

ipel

-0.73

isf

-0.72

ortal

-0.71

ammy

-0.67

opter

-0.65

isd

-0.65

ussen

-0.64

ikk

-0.63

ann

-0.63

imeo

-0.63

POSITIVE LOGITS

Ĥª

0.75

 outright

0.73

kus

0.73

 unanimously

0.70

ively

0.66

ELY

0.65

ulous

0.64

otiation

0.63

pelled

0.63

lect

0.63

Activations Density 0.025%

words related to rejection or refusal

instances of the word "reject" and its derivatives

No Comments

No Known Activations