INDEX

Explanations

phrases related to performing actions

oai_token-act-pair · gpt-3.5-turbo

frequent and common pronouns or articles in a text

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

 owing

-0.69

iating

-0.66

[*

-0.66

printf

-0.66

potion

-0.64

bank

-0.63

strument

-0.63

pport

-0.61

ranch

-0.61

omer

-0.59

POSITIVE LOGITS

 Yourself

0.85

 yourselves

0.85

 yourself

0.76

 Citiz

0.72

 Survivors

0.71

 Flavor

0.70

 Dice

0.66

 Blend

0.65

 Cosmos

0.65

 Arcane

0.64

Activations Density 0.312%

phrases related to performing actions

frequent and common pronouns or articles in a text

No Comments

No Known Activations