INDEX

Explanations

words related to physical rocks or rock-related actions

oai_token-act-pair · gpt-3.5-turbo

references to geological features

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

ersive

-0.80

EMA

-0.78

nec

-0.74

ntil

-0.73

nea

-0.73

elvet

-0.69

Wr

-0.68

present

-0.67

tm

-0.67

liction

-0.67

POSITIVE LOGITS

 rocks

1.21

lake

1.00

frog

0.90

 bould

0.87

bags

0.85

mith

0.82

 Rocks

0.81

 rock

0.80

 stones

0.79

 sculptures

0.78

Activations Density 0.011%

words related to physical rocks or rock-related actions

references to geological features

No Comments

No Known Activations