INDEX

Explanations

words related to inspection, approval, or acknowledgment

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "pat" in various contexts

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

vous

-0.74

perse

-0.72

ãĤ¼ãĤ¦ãĤ¹

-0.70

ource

-0.70

ESSION

-0.69

 Abyssal

-0.69

nesota

-0.67

actionDate

-0.67

ADE

-0.64

 Monstrous

-0.63

POSITIVE LOGITS

rolled

1.22

reon

1.22

ricia

1.21

hetically

1.19

terson

1.10

ric

1.08

ting

1.08

pat

1.02

tered

1.00

riot

1.00

Activations Density 0.010%

words related to inspection, approval, or acknowledgment

occurrences of the word "pat" in various contexts

No Comments

No Known Activations