INDEX

Explanations

phrases related to specific events or actions

oai_token-act-pair · gpt-3.5-turbo

references to significant events or items with numerical associations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ĸļ

-0.75

 Sturgeon

-0.70

epad

-0.64

bsite

-0.62

brance

-0.62

uria

-0.61

 Barkley

-0.61

erity

-0.61

itÃ©

-0.60

oshop

-0.60

POSITIVE LOGITS

 variants

1.09

 equivalents

1.07

 varieties

1.05

 iterations

1.04

 instances

1.04

isations

1.00

 implementations

1.00

 configurations

1.00

 endings

1.00

 fronts

1.00

Activations Density 0.857%

phrases related to specific events or actions

references to significant events or items with numerical associations

No Comments

No Known Activations