INDEX

Explanations

words related to directions or actions related to directionality

oai_token-act-pair · gpt-3.5-turbo

references to backwardness and forwardness, especially in the context of compatibility or progression

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ateurs

-0.79

 Parenthood

-0.79

anan

-0.79

tha

-0.76

ioxide

-0.74

ittee

-0.73

heon

-0.73

atum

-0.73

chens

-0.71

ateur

-0.71

POSITIVE LOGITS

 compatibility

1.03

robe

0.88

 glances

0.88

 causation

0.86

 slash

0.83

ward

0.82

 spiral

0.79

 glance

0.78

facing

0.77

 compat

0.76

Activations Density 0.044%

words related to directions or actions related to directionality

references to backwardness and forwardness, especially in the context of compatibility or progression

No Comments

No Known Activations