INDEX

Explanations

mentions or references to the word "human."

oai_token-act-pair · gpt-3.5-turbo

references to human characteristics and experiences

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

arella

-0.82

è¦

-0.81

armac

-0.76

OHN

-0.68

urations

-0.68

liga

-0.68

é¾įå

-0.67

abb

-0.67

effective

-0.66

kick

-0.65

POSITIVE LOGITS

 beings

1.34

itar

1.08

oids

1.07

itarian

1.02

 readable

0.99

istic

0.96

 embryonic

0.93

oid

0.92

ized

0.90

izing

0.88

Activations Density 0.026%

mentions or references to the word "human."

references to human characteristics and experiences

No Comments

No Known Activations