INDEX

Explanations

references to physical residences or dwellings

oai_token-act-pair · gpt-3.5-turbo

references to homes and residential structures

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

iasis

-0.68

 Airways

-0.63

BILITY

-0.63

pus

-0.61

asm

-0.61

 Delta

-0.61

 Wilde

-0.60

 constellation

-0.60

Magikarp

-0.60

Published

-0.58

POSITIVE LOGITS

chool

1.81

pun

1.07

creen

0.97

erver

0.95

ick

0.93

poons

0.91

cu

0.91

poon

0.90

nikov

0.89

ourced

0.87

Activations Density 0.030%

references to physical residences or dwellings

references to homes and residential structures

No Comments

No Known Activations