INDEX

Explanations

phrases related to balancing or counteracting something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the action of "output" or "producing results."

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

antry

-0.88

 confir

-0.68

Export

-0.61

 fundament

-0.60

 arsen

-0.59

 turnover

-0.59

 Primordial

-0.58

tyr

-0.58

oppy

-0.56

EY

-0.56

POSITIVE LOGITS

stretched

1.17

wards

1.14

posts

1.11

casts

1.06

fitted

1.06

flows

0.87

grown

0.85

skirts

0.85

fitting

0.83

ta

0.83

Activations Density 0.084%

phrases related to balancing or counteracting something

phrases indicating the action of "output" or "producing results."

No Comments

No Known Activations