INDEX

Explanations

words related to underground activities or locations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "under"

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

ãĥ£

-0.79

IER

-0.77

bernatorial

-0.75

 goodbye

-0.74

illac

-0.67

Tik

-0.67

-0.66

utenberg

-0.66

auga

-0.66

 Edison

-0.64

POSITIVE LOGITS

wear

1.09

lings

1.04

stood

1.01

ground

1.00

neath

0.96

graduate

0.96

lying

0.92

whelming

0.91

mentioned

0.90

lander

0.89

Activations Density 0.027%

words related to underground activities or locations

instances of the word "under"

No Comments

No Known Activations