INDEX

Explanations

words related to blades or cutting objects

oai_token-act-pair · gpt-3.5-turbo

references to different types of blades or blade-related terminology

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

ADRA

-0.86

quickShipAvailable

-0.83

bos

-0.82

gdala

-0.77

upon

-0.76

ESA

-0.75

riad

-0.74

ravel

-0.71

otional

-0.70

ally

-0.67

POSITIVE LOGITS

 blades

1.14

tips

1.03

 Runner

1.02

 blade

1.00

blade

0.89

tip

0.88

 wielded

0.82

 Cutter

0.81

hawk

0.81

 Blade

0.81

Activations Density 0.019%

words related to blades or cutting objects

references to different types of blades or blade-related terminology

No Comments

No Known Activations