INDEX

Explanations

keywords related to reacting or responding to various situations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "respond" with varying frequencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

fi

-0.71

 Marin

-0.69

 Blazers

-0.69

cutting

-0.68

rome

-0.68

colo

-0.68

 Demons

-0.68

ãĥĩãĤ£

-0.67

 Lions

-0.67

 Thieves

-0.67

POSITIVE LOGITS

 favorably

1.09

 positively

0.96

ivated

0.94

 harshly

0.86

ively

0.86

 adequately

0.85

 angrily

0.84

 appropriately

0.83

iments

0.83

 thereto

0.83

Activations Density 0.040%

keywords related to reacting or responding to various situations

instances of the word "respond" with varying frequencies

No Comments

No Known Activations