INDEX

Explanations

words related to professions or career paths

oai_token-act-pair · gpt-3.5-turbo

terms related to pro-activity or pro-situations 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-1.01

gow

-0.79

ãĤ¼ãĤ¦ãĤ¹

-0.79

 Dwell

-0.78

 Halls

-0.75

 Abyss

-0.71

 Sins

-0.71

 Grind

-0.69

 Seas

-0.69

 Leviathan

-0.67

POSITIVE LOGITS

digy

1.35

verbs

1.08

actively

0.91

 wrestler

0.89

oster

0.87

pro

0.86

scribed

0.85

ccess

0.82

ctor

0.81

zin

0.80

Activations Density 0.010%

words related to professions or career paths

terms related to pro-activity or pro-situations in various contexts

No Comments

No Known Activations