INDEX

Explanations

words related to personal names or business/organization names

oai_token-act-pair · gpt-3.5-turbo

the presence of names, possibly focusing on notable individuals or characters 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

URA

-0.70

ãĤ®

-0.69

ãĤ´

-0.68

ADRA

-0.65

 objective

-0.65

fid

-0.62

BIL

-0.62

 tremend

-0.61

 threat

-0.61

Premium

-0.60

POSITIVE LOGITS

aneers

0.92

hemy

0.85

reau

0.82

illet

0.78

Ã©n

0.77

 Ferry

0.77

doms

0.76

merce

0.76

estone

0.76

kees

0.74

Activations Density 0.038%

words related to personal names or business/organization names

the presence of names, possibly focusing on notable individuals or characters in various contexts

No Comments

No Known Activations