INDEX

Explanations

numerical quantities and measurements

oai_token-act-pair · gpt-3.5-turbo

numerical limits or capacities

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

Pear

-0.79

ãĥĺ

-0.77

 Kinnikuman

-0.68

ãĤ¢ãĥ«

-0.67

Reviewer

-0.67

 Lauder

-0.66

Rus

-0.66

Theme

-0.64

Chair

-0.64

 Guer

-0.63

POSITIVE LOGITS

 depending

0.86

dozen

0.82

depending

0.77

 percent

0.77

otine

0.73

 servings

0.73

uously

0.72

per

0.72

km

0.72

ousand

0.72

Activations Density 0.165%

numerical quantities and measurements

numerical limits or capacities

No Comments

No Known Activations