INDEX

Explanations

phrases related to connections or associations

oai_token-act-pair · gpt-3.5-turbo

phrases related to ties or connections 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 @ 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

perty

-0.84

issance

-0.80

ICAN

-0.74

sburg

-0.74

TPPStreamerBot

-0.71

ickr

-0.68

ister

-0.67

ocate

-0.66

isters

-0.66

================================================================

-0.66

POSITIVE LOGITS

 knot

1.16

breaker

1.06

breakers

0.97

tie

0.96

red

0.94

 knots

0.93

ged

0.88

dye

0.87

 breaker

0.87

breaking

0.83

Activations Density 0.050%

phrases related to connections or associations

phrases related to ties or connections in various contexts

No Comments

No Known Activations