INDEX

Explanations

names or terms ending with 'd'

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "did" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 constitu

-0.80

éĹĺ

-0.79

EStream

-0.79

 mosqu

-0.77

 Remem

-0.75

Nun

-0.72

 Pastebin

-0.71

OIL

-0.69

CLASSIFIED

-0.67

 suspense

-0.66

POSITIVE LOGITS

ynamic

1.27

rown

1.15

iamond

1.15

iary

1.10

ividual

1.09

etermin

1.09

etermined

1.08

orm

1.07

orf

1.05

olf

1.04

Activations Density 0.037%

names or terms ending with 'd'

references to the concept of "did" in various contexts

No Comments

No Known Activations