INDEX

Explanations

phrases related to superiority or advantage

oai_token-act-pair · gpt-3.5-turbo

references to something that is positioned at a higher level or status

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

tnc

-0.85

uality

-0.82

eln

-0.81

 Machina

-0.79

arium

-0.74

rx

-0.71

inatory

-0.69

printf

-0.68

UAL

-0.65

 Killer

-0.65

POSITIVE LOGITS

most

1.19

 bounds

0.81

 shaft

0.80

ĸļå£«

0.77

 tiers

0.77

 earners

0.75

lip

0.75

parts

0.75

ments

0.74

foundland

0.74

Activations Density 0.018%

phrases related to superiority or advantage

references to something that is positioned at a higher level or status

No Comments

No Known Activations