INDEX

Explanations

words related to actions or experiences

oai_token-act-pair · gpt-3.5-turbo

sentiments and expressions of loss or concern among people

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ertodd

-0.68

isms

-0.60

wise

-0.60

gery

-0.59

 grin

-0.59

 laughs

-0.58

 Garage

-0.58

pex

-0.57

 Medium

-0.56

 reluct

-0.56

POSITIVE LOGITS

 themselves

1.03

 selves

1.03

selves

0.99

atars

0.75

 outnumbered

0.73

 careers

0.72

 collectively

0.72

 THEIR

0.70

 quotas

0.70

 counterparts

0.69

Activations Density 0.507%

words related to actions or experiences

sentiments and expressions of loss or concern among people

No Comments

No Known Activations