INDEX

Explanations

words related to temporary or spontaneous events or establishments, often characterized by quick setup or appearance

oai_token-act-pair · gpt-3.5-turbo

references to pop-up events and culture

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

¿½

-0.71

 Norris

-0.70

 Cohn

-0.66

 Hyde

-0.64

 Howe

-0.64

Cth

-0.64

lly

-0.62

Dag

-0.62

©¶æ¥µ

-0.61

KC

-0.61

POSITIVE LOGITS

estate

1.00

favorite

0.98

spot

0.94

arts

0.93

weather

0.90

artist

0.90

trial

0.89

state

0.88

bucks

0.87

travel

0.85

Activations Density 0.066%

words related to temporary or spontaneous events or establishments, often characterized by quick setup or appearance

references to pop-up events and culture

No Comments

No Known Activations