INDEX

Explanations

references to concepts that are continually changing or increasing

oai_token-act-pair · gpt-3.5-turbo

references to the concept of continuous change or growth over time

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

seless

-0.71

RPG

-0.70

ItemTracker

-0.68

 Beir

-0.66

resy

-0.65

emia

-0.65

é¾

-0.64

 Legions

-0.64

 Tags

-0.63

Obj

-0.62

POSITIVE LOGITS

green

0.94

lasting

0.92

theless

0.89

THING

0.83

more

0.75

 dreamed

0.70

thing

0.69

entimes

0.69

icity

0.65

 wondered

0.64

Activations Density 0.021%

references to concepts that are continually changing or increasing

references to the concept of continuous change or growth over time

No Comments

No Known Activations