INDEX

Explanations

words related to negative emotions or situations, particularly focusing on the adjective "grim."

oai_token-act-pair · gpt-3.5-turbo

the word "grim" and its variations in different contexts

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

ertodd

-0.80

ournals

-0.75

adesh

-0.72

":[

-0.71

Ô

-0.70

BU

-0.70

Shares

-0.69

chnology

-0.66

aptic

-0.65

":[{"

-0.65

POSITIVE LOGITS

oire

1.11

acing

1.04

dark

1.00

ace

0.95

 grim

0.89

aces

0.87

 outlook

0.85

aced

0.83

lock

0.79

grim

0.79

Activations Density 0.027%

words related to negative emotions or situations, particularly focusing on the adjective "grim."

the word "grim" and its variations in different contexts

No Comments

No Known Activations