INDEX

Explanations

instances where something is being compared, analyzed, or represented

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reflect" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ccess

-0.71

Ã§a

-0.70

ciating

-0.66

Ñĭ

-0.65

load

-0.65

corn

-0.65

chief

-0.64

hattan

-0.63

GAN

-0.63

phans

-0.62

POSITIVE LOGITS

 reflect

1.06

 reflects

0.96

 reflections

0.96

 reflecting

0.95

 reflected

0.93

reflect

0.93

 reflection

0.90

ively

0.88

 Reflect

0.86

 sunlight

0.76

Activations Density 0.015%

instances where something is being compared, analyzed, or represented

instances of the word "reflect" in various contexts

No Comments

No Known Activations