INDEX

Explanations

words related to forceful and impactful actions

oai_token-act-pair · gpt-3.5-turbo

words related to flashy or attention-grabbing actions

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

ervation

-0.80

 Gutenberg

-0.77

yden

-0.76

otype

-0.75

brance

-0.74

swer

-0.74

asure

-0.73

elsen

-0.71

ussen

-0.71

communication

-0.71

POSITIVE LOGITS

OUT

0.80

Hur

0.77

IELD

0.76

arthy

0.76

ASH

0.75

hur

0.74

eed

0.73

Comment

0.73

Ùħ

0.72

UFF

0.70

Activations Density 0.020%

words related to forceful and impactful actions

words related to flashy or attention-grabbing actions

No Comments

No Known Activations