INDEX

Explanations

names or terms related to "wil" and "hel"

oai_token-act-pair · gpt-3.5-turbo

names and terms associated with specific individuals or entities

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

 juices

-0.69

ulo

-0.69

 Bucc

-0.65

 elig

-0.64

orrow

-0.62

ento

-0.61

 circulation

-0.60

entially

-0.60

uana

-0.60

livious

-0.59

POSITIVE LOGITS

stad

1.05

opol

0.84

ãĤ´ãĥ³

0.73

inki

0.73

phant

0.73

å§«

0.72

steen

0.71

gaard

0.71

Ń·

0.70

hurst

0.70

Activations Density 0.139%

names or terms related to "wil" and "hel"

names and terms associated with specific individuals or entities

No Comments

No Known Activations