INDEX

Explanations

researchers and institutions

oai_token-act-pair · gpt-3.5-turbo

instances of research affiliations or institutional references

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

ãĥĺ

-0.77

Quantity

-0.75

washer

-0.73

INGS

-0.72

ãĥĵ

-0.70

ãĤ¼ãĤ¦ãĤ¹

-0.68

ãĢĤ

-0.64

ãĥĸ

-0.64

INS

-0.63

 entit

-0.63

POSITIVE LOGITS

sembly

0.80

 alike

0.71

 Colleges

0.69

 allied

0.66

 convened

0.65

hip

0.64

hran

0.63

 POLITICO

0.63

vae

0.63

rak

0.61

Activations Density 0.293%

researchers and institutions

instances of research affiliations or institutional references

No Comments

No Known Activations