INDEX

Explanations

names of places or people

oai_token-act-pair · gpt-3.5-turbo

specific sequences of letters or patterns, likely tied to names or identifiers

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Pastebin

-0.89

 Dropbox

-0.67

 blender

-0.64

 Gutenberg

-0.61

 Razer

-0.61

 Hilbert

-0.60

 Blaz

-0.60

 Cortex

-0.59

NPR

-0.58

 stim

-0.58

POSITIVE LOGITS

ãĤ¨ãĥ«

0.92

hesion

0.80

ativity

0.70

entials

0.69

facing

0.69

ath

0.69

theless

0.68

erest

0.68

/-

0.68

apest

0.68

Activations Density 0.300%

names of places or people

specific sequences of letters or patterns, likely tied to names or identifiers

No Comments

No Known Activations