INDEX

Explanations

references to tunnels.

oai_token-act-pair · gpt-3.5-turbo

references to tunnels and their 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

itia

-0.76

CHO

-0.74

ģĸ

-0.65

 Liberties

-0.65

Vict

-0.65

orian

-0.65

ãĤŃ

-0.64

rament

-0.63

oln

-0.63

 ANGEL

-0.62

POSITIVE LOGITS

 tunnel

1.07

 tunnels

0.98

 boring

0.90

ibur

0.90

 Tunnel

0.86

dug

0.82

vision

0.81

 drilled

0.79

dump

0.78

 handshake

0.77

Activations Density 0.009%

references to tunnels.

references to tunnels and their various contexts

No Comments

No Known Activations