INDEX

Explanations

terms related to physical locations or conditions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "par" used 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ħĭ

-1.00

xia

-0.92

æĸ¹

-0.74

hower

-0.72

éĹĺ

-0.71

hirt

-0.69

UTH

-0.68

sonian

-0.66

¥µ

-0.66

BILITIES

-0.65

POSITIVE LOGITS

par

0.91

par

0.88

onto

0.82

allel

0.81

rot

0.80

icularly

0.79

pend

0.77

rots

0.75

ross

0.74

 ILCS

0.73

Activations Density 0.006%

terms related to physical locations or conditions

instances of the word "par" used in various contexts

No Comments

No Known Activations